Enzo Ferrante edited untitled.tex  almost 8 years ago

Commit id: 9995d35ea7aa6b815866fba5a57e168f2b69dd70

deletions | additions      

       

\textbf{ Abstract}  The main goal of the this  project is to extract high level semantic data from soccer matchesgiven only the  video of a match and sequences,  using Machine Learning machine learning  and Computer Vision computer vision  techniques. \begin{itemize}  \item During the first year year,  Ihave  explored the available state-of-the-art  techniques allowing that could be used  to extract the players' player  trajectories given the videos of a match with several static camera camera.  \item During the second year year,  I focused on the problem of player trajectory  clustering of the players trajectories which leads to the thorny issue of the choice of problem. I particularly focused on designing  a reliable and efficient metric to establish correspondences  between trajectories. \end{itemize}